Interactive plans and reports in your AI code editor.
Research should not arrive as a wall of chat. In RevampKey the researcher hands back a report you can sort, filter and question, and larger changes become plans split into steps you can review before any code is written.
A report you can work with, not just read.
This one audits an appointments app before self-serve booking is added. Filter it by severity. In the app, you can also sort tables, follow a claim to its source and ask for a revision.
Booking readiness audit
What the appointments app handles today, and what self-serve booking needs first.
Four of the nine booking paths have no handler today.
Two clients can confirm the same time at the same momentNo lock on a time slot today
A reschedule can leave the old booking behindThe old time is never released
A held time never frees if the browser closesHolds have no expiry
Reminders can go out at the wrong hour across time zonesTime is stored without a zone
A cancel inside 24 hours is not flagged to the front deskNeeds your decision
The confirmation has no calendar attachmentNice to have
Findings you can check and act on.
Sort and filter
Tables respond to you. Find the high-risk items first instead of scrolling a long reply.
See the source
Each claim points back to where it came from, in your project or on the web, so you can trust it or push back.
Ask for a revision
Say what is missing in plain words. You get a new revision and the earlier ones stay for comparison.
Large changes become a plan you can review in steps.
When a request needs more than a quick edit, RevampKey shows a plan before it builds. The plan is divided into modules, each with a goal, an area it touches and a checklist, so a missing dependency or a scope decision shows up while it is still cheap to change.
Planning adapts to the request: short tasks go straight to review. Details are in the docs on plans and modules.
Plans and reports live in your project.
Files, not a service
Specs, trackers and diagrams are files in your repository with revision history. Nothing is stored on our servers.
Decisions are recorded
Your answers to questions are written into the plan, so the reasoning outlasts the chat.
Reuse earlier work
Past reports can inform new tasks, so the same investigation is not run twice.
See also how earlier reports are reused.
Plans and reports, answered.
What is an interactive report?
A report the researcher agent builds from your project and the web, shown as a real document instead of a long chat reply. You can sort and filter tables, see where each claim came from, and ask for a change.
What happens when I ask for a change to a report?
You ask in plain words and a new revision is produced. Earlier revisions stay available, so you can compare what changed.
Does every task get a plan?
No. Planning is a review surface, not a gate. A short task goes straight to review, while a larger change shows a plan with modules so the approach can be checked first. Plans and modules
Where are plans and reports stored?
In your own project, as files with revision history. Nothing is stored on RevampKey servers.
Can earlier reports be reused?
Yes. Past work reports can be offered as context so an investigation is not repeated, and you control whether main work and delegated work can see them. Past work reports
Keep exploring
How it works
Follow one idea through the researcher, product manager, designer, engineer and QA.
See the whole loop →Start with one idea.
Describe something you have been putting off. Read the plan, approve the screens, and decide whether to let it build.
macOS 12+ · Windows 10+ · Ubuntu 20.04+ · See what shipped in 2.8.0